#
#asahi-dev · 2024-10-07 06:59
nickchanthe new way to split device trees should ease adding more nodes in the future
#asahi-dev · 2024-10-07 06:58
nickchannow the a7-a11 pmgr stuff is figured out downstream I will give the device tree series another spin today
#asahi-dev · 2024-10-06 23:04
chadmedoh neat that was fast. thanks!
#asahi-dev · 2024-10-06 22:00
flokliok ty
#asahi-dev · 2024-10-06 20:10
jannauit include the ring sense irq work (and my patch to increase the debounce time which I forgot to drop)
#asahi-dev · 2024-10-06 20:09
floklichadmed: jannau: that tag includes the simple msleep commit only, not the later rework, right?
#asahi-dev · 2024-10-06 18:16
mpsthanks
#asahi-dev · 2024-10-06 18:10
jannauI've pushed asahi-6.11.2-1 to github
#asahi-dev · 2024-10-05 06:46
chadmedokay handling ring sense sanely lets us get rid of this issue while keeping the minimum debounce and no added delay
#asahi-dev · 2024-10-05 01:33
amarioguy(well for USB2 anyways)
#asahi-dev · 2024-10-05 00:36
amarioguynever mind lol, found the original commit for it, seems like iBoot does the PHY config, and the u-boot driver just resets the DWC3, so that's fine lol
#asahi-dev · 2024-10-05 00:33
amarioguyjust asking mainly because the ATC code in u-boot looks quite barren so just clarifying
#asahi-dev · 2024-10-05 00:04
amarioguystate of the controllers *when u-boot starts i mean, not after bringup
#asahi-dev · 2024-10-05 00:03
amarioguyquick question, when exiting m1n1 and chainloading into u-boot what's the state of the type-c controllers? do i need to bring up both the ATC PHY and the DWC3 (along with the PHY tunables) or is the PHY already taken care of in m1n1?
#asahi-dev · 2024-10-04 12:29
chadmedeither way i think smartly using ring sense fixes all possible issues so ill fix that up tomorrow
#asahi-dev · 2024-10-04 12:28
chadmedi reliably reproduce this by holding the plug in just enough to hit the tip sense switch in the jack but not enough to terminate all of the rings, then plugging it in all the way after plug event shows up in dmesg
#asahi-dev · 2024-10-04 12:27
chadmedthe issue is more the detection happening before the plug is fully seated. ive not once had a misdetection on a fully seated plug
#asahi-dev · 2024-10-04 12:26
jannauI find it a little unlikely that the 500ms debounce is insufficient 5 insertions in a row
#asahi-dev · 2024-10-04 12:24
jannauI'm not 100% convinced it's the jack insertion though. could be the mic detection setting with specific mics as well
#asahi-dev · 2024-10-04 12:21
chadmedi wonder if they suffer from the same issue on devices with dodgy jacks then
#asahi-dev · 2024-10-04 12:20
jannaujust unmask the irq after the full tip sense. on the other hand none of the other 2 cs42* driver's I looked at uses the ring sense irq
#asahi-dev · 2024-10-04 12:15
chadmedbut other than that ignore it, it's quite noisy
#asahi-dev · 2024-10-04 12:15
chadmedso if tip sense detects a partial plug, ring sense triggers a plug event which triggers a re-detection
#asahi-dev · 2024-10-04 12:14
chadmedim just playing with the best way to handle it so it's completely foolproof. im thinking we handle ring sense IFF we're already in a plugged state
#asahi-dev · 2024-10-04 12:14
chadmedyeah i added 1500ms for testing locally
#asahi-dev · 2024-10-04 12:13
jannaunot sure if you saw, I did for now only https://github.com/AsahiLinux/linux/commit/960018148bf7ac93f6832e4363211b9499310871
#asahi-dev · 2024-10-04 12:13
chadmedyeah
#asahi-dev · 2024-10-04 12:12
jannauthe tip handling is more an issue of the plug design and not of the wiring
#asahi-dev · 2024-10-04 12:12
chadmedbut the benefit is that it retriggers when you fully seat the plug, so you can re-run detection
#asahi-dev · 2024-10-04 12:11
chadmedit does trigger a plug event on a partial plug though
#asahi-dev · 2024-10-04 12:11
chadmedyeah i just unmasked it and its fine
#asahi-dev · 2024-10-04 12:10
jannauI don't think the ring sense irq is useless. it triggers only for headsets when the the jack is fully seated
#asahi-dev · 2024-10-04 12:10
chadmedoh never mind the interrupts were masked
#asahi-dev · 2024-10-04 11:50
chadmedi think the best we can do here is just max out the tip sense irq debounce and tell people theyre holding it wrong if they keep running into this
#asahi-dev · 2024-10-04 11:50
chadmedhmph the ring sense irq is useless, i think the issue is the jacks themselves or how theyre wired up and not the codec
#asahi-dev · 2024-10-03 22:13
chadmedthe easiest way woud be to patch the installer and let it do all the hard work for you
#asahi-dev · 2024-10-03 20:27
calebccffo/ hey all, how would one get set up to start working on m1n1 for M3? It seems like a bit of a chicken and egg problem to set up a new OS container, but I guess I'm missing something obvious
#asahi-dev · 2024-10-02 16:47
nickchanit is uart6
#asahi-dev · 2024-10-02 16:47
nickchanApple TV HD debug console is *not* uart0
#asahi-dev · 2024-10-02 16:47
nickchanfun
#asahi-dev · 2024-10-02 06:40
nickchanwouldn't want to declare it until something else in the dt needs it
#asahi-dev · 2024-10-02 06:40
nickchantoo cursed to declare it now
#asahi-dev · 2024-10-02 06:40
nickchanam also thinking of stripping cpu-map from the A11 device tree
#asahi-dev · 2024-10-01 23:37
chadmed_and then as you say include the soc dtsi directly for the apple tv
#asahi-dev · 2024-10-01 23:37
chadmed_nickchan: you could just do like txxxx-idevice.dtsi
#asahi-dev · 2024-10-01 17:09
nickchanso it will include the base SoC dtsi directly, anyways
#asahi-dev · 2024-10-01 17:09
nickchanand there is only one Apple TV for the SoC that have it
#asahi-dev · 2024-10-01 17:09
nickchanbut then it does not matter since apple tv is defintiely not an handheld
#asahi-dev · 2024-10-01 17:09
chaos_princesskinda? i think the name for the device class is "set-top box"
#asahi-dev · 2024-10-01 17:08
nickchanchaos_princess: is apple tv a desktop ?

plus ›